Professional Documents
Culture Documents
Mapy
Mapy
June 2010
Preface
Mathematics, rightly viewed, possesses not only truth, but supreme beauty a beauty cold
and austere, like that of sculpture, without appeal to any part of our weaker nature, without the
gorgeous trappings of painting or music, yet sublimely pure, and capable of a stern perfection
such as only the greatest art can show, wrote Bertrand Russell about the beauty of mathematics.
All of us may not reach such higher planes, probably reserved for Russels and Ramanujans, but
we also have beautiful curves and nice geometrical gures with intricate symmetries, like fractals,
generated by seemingly dull equations. This book attempts to explore it using a simple tool, the
Python programming language.
I started using Python for the Phoenix project (www.iuac.res.in). Phoenix was driven in to
Python by Pramode CE (pramode.net) and I followed. Writing this document was triggered by
some of my friends who are teaching mathematics at Calicut University.
In the rst chapter, a general introduction about computers and high level programming languages is given. Basics of Python language, Python modules for array and matrix manipulation,
2D and 3D data visualization, type-setting mathematical equations using latex and numerical
methods in Python are covered in the subsequent chapters. Example programs are given for every
topic discussed. This document is meant for those who want to try out these examples and modify
them for better understanding. Huge amount of material is already available online on the topics
covered, and the references to many resources on the Internet are given for the benet of the serious
reader.
This book comes with a live CD, containing a modied version of Ubuntu GNU/Linux operating
system. You can boot any PC from this CD and practice Python. Click on the 'Learn by Coding'
icon on the desktop to browse through a collection of Python programs, run any of them with
a single click.
You can practice Python very easily by modifying and running these example
programs.
Contents
1 Introduction
1.1
Hardware Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2
Software components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.1
1.2.2
1.2.2.1
. . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.3
The File-system
1.2.3.1
1.2.3.2
Current Directory
. . . . . . . . . . . . . . . . . . . . . . . . . . .
10
1.3
Text Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
1.4
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
1.5
On Free Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
1.6
Exercises
11
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 Programming in Python
2.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
. . . . . . . . . . . . . . . . . . . .
13
2.2
14
2.3
16
2.4
Python Strings
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
Slicing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
2.4.1
13
2.5
Python Lists
2.6
2.7
2.8
2.9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
2.8.1
20
2.8.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
21
. . . . . . . . . . . . . . . . . . . . . . . . . . .
22
23
2.12 Exercises
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
2.13 Functions
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
26
. . . . . . . . . . . . . . . . . . . . . . . .
27
27
28
28
28
CONTENTS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
29
2.15.2 Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
32
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
33
34
36
. . . . . . . . . . . . . . . . . . . . . . .
38
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
40
. . . . . . . . . . . . . . . . . . . . . . . . . .
40
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
44
44
45
3.1.1.1
45
3.1.1.2
. . . . . . . . . . . . . .
45
3.1.1.3
zeros(shape, datatype)
. . . . . . . . . . . . . . . . . . . . . . . .
45
3.1.1.4
ones(shape, datatype) . . . . . . . . . . . . . . . . . . . . . . . . .
45
3.1.1.5
random.random(shape)
. . . . . . . . . . . . . . . . . . . . . . . .
46
3.1.1.6
reshape(array, newshape) . . . . . . . . . . . . . . . . . . . . . . .
46
3.1.2
Copying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
3.1.3
Arithmetic Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
3.1.4
cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
3.1.5
dot product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
3.1.6
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
3.1.7
Matrix inversion
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
3.2
Vectorized Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
3.3
Exercises
49
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Data visualization
4.1
4.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
51
4.1.1
Multiple plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
52
4.1.2
Polar plots
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
4.1.3
Pie Charts
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
4.2.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
4.2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
4.2.3
Parametric plots
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
Famous Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
4.3.1
Astroid
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
4.3.2
Ellipse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
4.3.3
. . . . . . . . . . . . . . . . . . . . . . .
58
4.3.4
Polar Rose
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
4.4
Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
4.5
Fourier Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
60
4.6
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
61
4.7
Fractals
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
62
4.3
CONTENTS
4.8
Meshgrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
4.9
3D Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
4.9.1
Surface Plots
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
4.9.2
Line Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
4.9.3
Wire-frame Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
65
4.11 Exercises
66
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
68
5.1
Document classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2
Modifying Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
5.3
69
5.4
Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
70
5.5
Typesetting Equations
71
5.5.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
68
. . . . . . . . . . . . . . . . . . .
72
5.6
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
73
5.7
74
5.8
Example Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
5.9
Exercises
76
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6 Numerical methods
6.1
Derivative of a function
6.1.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
78
79
6.2
Numerical Integration
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
80
6.3
82
6.3.1
Euler method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
6.3.2
Runge-Kutta method
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
6.3.3
85
Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
86
6.4.1
Taylor's Series
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
6.4.2
89
91
6.5.1
Method of Bisection
92
6.5.2
Newton-Raphson Method
6.4
6.5
6.6
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
93
95
6.6.1
. . . . . . . . . . . . . . . . . . . .
95
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
97
6.7
6.8
Interpolation
6.8.1
6.9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Newton's polynomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Exercises
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . .
97
101
102
102
105
108
. . . . . . . . . . . . . . . . . . . . .
108
108
Chapter 1
Introduction
Primary objective of this book is to explore the possibilities of using Python language as a tool
for learning mathematics and science. The reader is not assumed to be familiar with computer
programming. Ability to think logically is enough. Before getting into Python programming, we
will briey explain some basic concepts and tools required.
Computer is essentially an electronic device like a radio or a television. What makes it dierent
from a radio or a TV is its ability to perform dierent kinds of tasks using the same electronic
and mechanical components. This is achieved by making the electronic circuits exible enough to
work according to a set of instructions. The electronic and mechanical parts of a computer are
called the Hardware and the set of instructions is called Software (or computer program). Just by
changing the Software, computer can perform vastly dierent kind of tasks. The instructions are
stored in binary format using electronic switches.
and an Arithmetic and Logic Unit, ALU. The control unit brings the instructions stored in the
main memory one by one and acts according to it.
tween memory and input/output units. The ALU can perform arithmetic operations like addition,
multiplication and logical operations like comparing two numbers.
Memory stores the instructions and data, that is processed by the CPU. All types of information
are stored as binary numbers. The smallest unit of memory is called a binary digit or Bit. It can
have a value of zero or one. A group of eight bits are called a Byte. A computer has Main and
Secondary types of memory.
memory.
Before processing, data and instructions are moved into the main
location by using it's address. Main memory is made of semiconductor switches and is very fast.
There are two types of Main Memory. Read Only Memory and Read/Write Memory. Read/Write
Memory is also called Random Access Memory. All computers contains some programs in ROM
which start running when you switch on the machine. Data and programs to be stored for future
use are saved to Secondary memory, mainly devices like Hard disks, oppy disks, CDROM or
magnetic tapes.
1 The cabinet that encloses most of the hardware is called CPU by some, mainly the computer vendors. They
are not referring to the actual CPU chip.
CHAPTER 1.
INTRODUCTION
The Input devices are for feeding the input data into the computer.
common input device. Mouse, scanner etc. are other input devices. The processed data is displayed
or printed using the output devices. The monitor screen and printer are the most common output
devices.
in control. On user request, operating system loads other application programs from disk to the
main memory and executes them. OS also provides a le system, a facility to store information on
devices like oppy disk and hard disk. In fact the OS is responsible for managing all the hardware
resources.
GNU/Linux and MS Windows are two popular operating systems. Based on certain features,
operating systems can be classied as:
Single user, single process systems like MS DOS. Only one process can run at a time. Such
operating systems do not have much control over the application programs.
Multi-tasking systems like MS Windows, where more than one processe can run at a time.
Multi-user, multi-tasking systems like GNU/Linux, Unix etc. More than one person can use
the computer at the same time.
Real-time systems, mostly used in control applications, where the response time to any
external input is maintained under specied limits.
Under a
Graphical User Interface, we do this by choosing the desired application from a menu.
possible only because someone has added it to the menu earlier.
It is
own programs, obviously they will not appear in any menu. Another way to request the operating
system to execute a program is to enter the name of the program (more precisely, the name of
CHAPTER 1.
INTRODUCTION
start a Terminal from the menu names Applications->Accessories->Terminal. Figure 1.1 shows a
Terminal displaying the list of les in a directory (output of the command 'ls -l' , the -l option is
for long listing).
The command processor oers a host of features to make the interaction more comfortable. It
keeps track of the history of commands and we can recall previous commands, modify and reuse
them using the cursor keys. There is also a completion feature implemented using the Tab key
that reduces typing. Use the tab key to complete command and lenames. To run
our test directory, type
python h
hello.py
from
and then press the tab key to complete it. If there are more than
one le starting with 'h', you need to type more characters until the ambiguity is removed. Always
use the up-cursor key to recall the previous commands and re-issue it.
The commands given at the terminal are processed by a program called the
shell.
(The version
now popular under GNU/Linux is called bash, the Bourne again shell). Some of the GNU/Linux
commands are listed below.
top : Shows the CPU and memory usage of all the processes started.
cp lename lename : copies a le to another.
mv : moves les from one folder to another, or rename a le.
rm : deletes les or directories.
man : display manual pages for a program. For example 'man bash' will give help on the
bash shell. Press 'q' to come out of the help screen.
See the manual pages of 'mv', cp, 'rm' etc. to know more about them. Most of these commands
are application programs, stored inside the folders /bin or /sbin, that the shell starts for you and
displays their output inside the terminal window.
CHAPTER 1.
INTRODUCTION
root
group
who has permission to use all the resources. Ordinary user accounts, with a username and password,
are created for everyone who wants to use the computer. In a multi-user operating system, like
GNU/Linux, every user will have one directory inside which he can create sub-directories and les.
This is called the 'home directory' of that user. Home directory of one user cannot be modied by
another user.
The operating system les are owned by
root.
fred
/home/fred
(g
1.2) and its contents. That is also called the user's home directory. Every le and directory has
three types of permissions : read, write and execute. To view them use the 'ls -l ' command. The
rst character of output line tells the type of the le.
rwx
(read, write, execute) permissions for the owner of that le. Next three for the users belonging to
the same group and the next three for other users. A hyphen character (-) means the permission
corresponding to that eld is not granted. For example, the gure 1.1 shows a listing of ve les:
1. asecret.dat : read & write for the owner. No one else can even see it.
2. foo.png : rw for owner, but others can view the le.
3. hello.py : rwx for owner, others can view and execute.
4. share.tex : rw for owner and other members of the same group.
CHAPTER 1.
10
INTRODUCTION
5. xdata is a directory. Note that execute permission is required to view contents of a directory.
The system of ownerships and permissions also protects the system from virus attacks .
The
virus programs damage the system by modifying some application program. On a true multi-user
system, for example GNU/Linux, the application program and other system les are owned by
the root user and ordinary users have no permission to modify them. When a virus attempts to
modify an application, it fails due to this permission and ownership scheme.
'gedit'
Using them one can create a program without knowing much about the computer
hardware.
We already learned that to solve a problem we require an algorithm and it has to be executed
step by step.
A high level
programming language is a set of well dened notations which is capable of expressing algorithms.
notations.
2 Do
not expect this from the MS Windows system. Even though it allows to create users, any user ( by running
programs or viruses) is allowed to modify the system les. This may be because it grew from a single process system
like MSDOS and still keeps that legacy.
CHAPTER 1.
11
INTRODUCTION
4. A set of syntax rules that precisely specify the combination of words and symbols permissible
in the language.
5. A set of semantic rules that assigns a single, precise and unambiguous meaning to each
syntactically correct statement.
Program text written in a high level language is often called the Source Code. It is then translated into the machine language by using translator programs. There are two types of translator
High level languages make the programming job easier. We can write programs that are machine
independent. For the same program dierent compilers can produce machine language code to run
on dierent types of computers and operating systems. BASIC, COBOL, FORTRAN, C, C++,
Python etc.
are some of the popular high level languages, each of them having advantages in
dierent elds.
To write any useful program for solving a problem, one has to develop an algorithm.
The
notations and rules of the language one can start writing programs to implement more complicated
algorithms.
software, the human-readable form of the program (the source code) must be made available to
the recipient along with a notice granting the above permissions.
The free software movement was conceived in 1983 by Richard Stallman to give the benet of
"software freedom" to computer users. Stallman founded the Free Software Foundation in 1985
to provide the organizational structure to advance his Free Software ideas. Later on, alternative
movements like Open Source Software came.
Software for almost all applications is currently available under the pool of Free Software.
1.6 Exercises
1. What are the basic hardware components of a computer.
2. Name the working directory of a user named 'ramu' under GNU/Linux.
3. What is the command to list the le names inside a directory (folder).
4. What is the command under GNU/Linux to create a new folder.
5. What is the command to change the working directory.
CHAPTER 1.
INTRODUCTION
6. Can we install more than one operating systems on a single hard disk.
7. Name two most popular Desktop Environments for GNU/Linux.
8. How to open a command window from the main menu of Ubuntu GNU/Linux.
9. Explain the le ownership and permission scheme of GNU/Linux.
12
Chapter 2
Programming in Python
Python is a simple, high level language with a clean syntax. It oers strong support for integration
with other languages and tools, comes with extensive standard libraries, and can be learned in a
few days. Many Python programmers report substantial productivity gains and feel the language
encourages the development of higher quality, more maintainable code. To know more visit the
Python website.
Text Editor
and the
Terminal
from the
Applications->Accessories menu.
this method only for viewing the results of single Python statements, for example to use Python
as a calculator. It could be confusing when you start writing larger programs, having looping and
conditional statements. The preferred way is to enter your source code in a text editor, save it to
a le (with .py extension) and execute it from the command terminal using Python. A screen-shot
of the Desktop with Text Editor and Terminal is shown in gure 2.1.
In this document, we will start writing small programs showing the essential elements of the
language without going into the details. The reader is expected to run these example programs and
also to modify them in dierent ways.It is like learning to drive a car, you master it by practicing.
Let us start with a program to display the words
Hello World
the customary 'hello world' program. There is another version that prints 'Goodbye cruel world',
probably invented by those who give up at this point. The Python 'hello world' program is shown
below.
1 http://www.python.org/
http://docs.python.org/tutorial/
This document, example programs and a GUI program to browse through them are at
http://www.iuac.res.in/phoenix
13
CHAPTER 2.
14
PROGRAMMING IN PYTHON
Example. hello.py
print 'Hello World'
This should be entered into a text le using any text editor. On a GNU/Linux system you may
use the text editor like 'gedit' to create the source le, save it as
hello.py .
the Python Interpreter to execute the new program. For that, open a command terminal and (at
the $ prompt) type:
$ python hello.py
y = f loat(3).
rst.py
shows how to dene variables of dierent data types. It also shows how
2 For quick practicing, boot from the CD provided with this book and click on the learn-by-coding icon to browse
through the example programs given in this book. The browser allows you to run any of them with a single click,
modify and save the modied versions.
CHAPTER 2.
15
PROGRAMMING IN PYTHON
Example: rst.py
'
A multi-line comment, within a pair of three single quotes.
In a line, anything after a # sign is also a comment
'
x = 10
print x, type(x)
# print x and its type
x = 10.4
print x, type(x)
x = 3 + 4j
print x, type(x)
x = 'I am a String '
print x, type(x)
The output of the program is shown below. Note that the type of the variable
changes during
10 <type 'int'>
10.4 <type 'float'>
(3+4j) <type 'complex'>
I am a String <type 'str'>
The program treats the variables like humans treat labelled envelopes. We can pick an envelope,
write some name on it and keep something inside it for future use. In a similar manner the program
creates a variable, gives it a name and keeps some value inside it, to be used in subsequent steps.
So far we have used four data types of Python: int, oat, complex and str. To become familiar with
them, you may write simple programs performing arithmetic and logical operations using them.
Example: oper.py
x = 2
y = 4
print x + y * 2
s = 'Hello '
print s + s
print 3 * s
print x == y
print y == 2 * x
Running the program
oper.py
10
Hello Hello
Hello Hello Hello
False
True
Note that a String can be added to another string and it can be multiplied by an integer. Try to
understand the logic behind that and also try adding a String to an Integer to see what is the error
message you will get. We have used the logical operator
==
CHAPTER 2.
16
PROGRAMMING IN PYTHON
Operator
or
and
not x
in, not in
<, <=, >, >=, !=, ==
|
^
&
< <, > >
+,*, /, %
+x , -x
~
**
x[index]
Description
Boolean OR
Boolean AND
Boolean NOT
Membership tests
Comparisons
Bitwise OR
Bitwise XOR
Bitwise AND
Bitwise Shifting
Add, Subtract
Multiply, divide, reminder
Positive, Negative
Bitwise NOT
Exponentiation
Subscription
Expression
0 or 4
3 and 0
not 0
3 in [2.2,3,12]
2>3
1|2
1^5
1&3
1 << 3
6-4
5%2
-5*2
~1
2 ** 3
a='abcd' ; a[1]
Result
4
0
True
True
False
3
4
1
8
2
1
-10
-2
8
'b'
2+3*4,
is the
addition done rst or the multiplication? According to elementary arithmetics, the multiplication
should be done rst.
It means that the multiplication operator has higher precedence than the
addition operator. If you want the addition to be done rst, enforce it by using parenthesis like
(2 + 3) 4.
evaluation.
Example: string.py
s = 'hello world'
print s[0]
print s[1]
print s[-1]
Example: string2.py
a = 'hello'+'world'
print a
b = 'ha' * 3
string.py.
CHAPTER 2.
17
PROGRAMMING IN PYTHON
print b
print a[-1] + b[0]
will give the output
helloworld
hahaha
dh
The last element of
2.4.1 Slicing
Part of a String can be extracted using the slicing operation. It can be considered as a modied
form of indexing a single character. Indexing using
s[a : b]
extracts elements
s[a]
to
s[b 1].
We
can skip one of the indices. If the index on the left side of the colon is skipped, slicing starts from
the rst element and if the index on right side is skipped, slicing ends with the last element.
Example: slice.py
a = 'hello world'
print a[3:5]
print a[6:]
print a[:5]
The reader can guess the nature of slicing operation from the output of this code, shown below.
'lo'
'world'
'hello'
Please note that specifying a right side index more than the length of the string is equivalent to
skipping it. Modify
slice.py
a[6 : 20]
to demonstrate it.
list1.py
elements.
Example: list1.py
a = [2.3, 3.5, 234] # make a list
print a[0]
a[1] = 'haha'
# Change an element
print a
The output is shown below
3.
3 The oating point number 2.3 showing as 2.2999999999999998 is interesting. This is the very nature of oting point representation of numbers, nothing to do with Python. With the precision we are using, the error in
representing 2.3 is around 2.0e-16. This becomes a concern in operations like inversion of big matrices.
CHAPTER 2.
18
PROGRAMMING IN PYTHON
2.3
[2.2999999999999998, 'haha', 234]
Lists can be sliced in a manner similar to that if Strings. List addition and multiplication are
demonstrated by the following example. We can also have another list as an element of a list.
Example: list2.py
a = [1,2]
print a * 2
print a + [3,4]
b = [10, 20, a]
print b
The output of this program is shown below.
[1, 2, 1, 2]
[1, 2, 3, 4]
[10, 20, [1, 2] ]
We can change the value of an element in a list, add new elements to it and remove any existing
element. This is not possible with String type. Uncomment the last line of third.py and run it to
clarify this point.
Example: third.py
s = [3, 3.5, 234]
s[2] = 'haha'
print s
x = 'myname'
#x[1] = 2
# make a list
# Change an element
# String type
# uncomment to get ERROR
The List data type is very exible, an element of a list can be another list. We will be using lists
extensively in the coming chapters. Tuple is another data type similar to List, except that it is
immutable. List is dened inside square brackets, tuple is dened in a similar manner but inside
parenthesis, like
raw_input()
input()
for numeric
Example: kin1.py
4 Functions will be introduced later. For the time being, understand that it is an isolated piece of code, called
from the main program with some input arguments and returns some output.
CHAPTER 2.
19
PROGRAMMING IN PYTHON
integer
or
oat
String
type
by 2 before printing. Multiplying a String by 2 prints it twice. If the String contains any other
characters than
Example: kin2.py
x,y = input('Enter x and y separated by comma ')
print 'The sum is ', x + y
s = raw_input('Enter a decimal number ')
a = float(s)
print s * 2
# prints string twice
print a * 2
# converted value times 2
We have learned about the basic data types of Python and how to get input data from the keyboard.
This is enough to try some simple problems and algorithms to solve them.
Example:
area.py
pi = 3.1416
r = input('Enter Radius ')
a = pi * r ** 2
# A = r2
print 'Area = ', a
r2 using the exponentiation
2
operator , and multiply it with using the multiplication operator . r is evaluated rst because
2
** has higher precedence than *, otherwise the result would be (r) .
The above example calculates the area of a circle. Line three calculates
Example:
badtable.py
print
print
print
print
print
1
2
3
4
5
*
*
*
*
*
8
8
8
8
8
CHAPTER 2.
20
PROGRAMMING IN PYTHON
Well, we are stopping here and looking for a better way to do this job.
The solution is to use the
while
loop of Python.
while
is evaluated, and if it is True, the body of the while loop (the indented lines below the while
statement) is executed.
We should
have some statement inside the body of the loop that will make this condition false after few
iterations. Otherwise the program will run in an innite loop and you will have to press Control-C
to terminate it.
The program
Example:
table.py
x = 1
while x <= 10:
print x * 8
x = x + 1
As per the Python syntax, the while statement ends with a colon and the code inside the
loop is indented.
while
accepted as forms of indentation in Python. Mixing spaces and tabs can create bugs that are hard
to nd, since the text editor does not show the dierence. There should not be any extra white
spaces in the beginning of any line.
The line before any indented block must end with a colon character.
for
loops are slightly dierent from the for loops of other languages.
Python
for
loop
iterates over a compound data type like a String, List or Tuple. During each iteration, one member
of the compound data is assigned to the loop variable. The exibility of this can be seen from the
examples below.
Example:
forloop.py
a = 'my name'
for ch in a: # ch is the loop variable
print ch
b = ['hello', 3.4, 2345, 3+5j]
for item in b:
print item
For constructing for loops that executes a xed number of times, we can create a list using the
range() function and run the for loop over that.
CHAPTER 2.
Example:
21
PROGRAMMING IN PYTHON
forloop2.py
mylist = range(5)
print mylist
for item in mylist:
print item
The range function in the above example will generate the list
[0, 1, 2, 3, 4].
It is possible to specify
the starting point and increment as arguments in the form range(start, end+1, step). The following
example prints the table of 5 using this feature.
Example:
forloop3.py
mylist = range(5,51,5)
for item in mylist:
print item
In some cases, we may need to traverse the list to modify some or all of the elements. This can be
done by looping over a list of indices generated by the range() function.For example, the program
forloop4.py zeros all the elements of the list.
Example:
forloop4.py
a = [2, 5, 3, 4, 12]
size = len(a)
for k in range(size):
a[k] = 0
print a
Example:
if
if, elif
compare.py
Example:
big.py
and
CHAPTER 2.
22
PROGRAMMING IN PYTHON
x > 10 and x < 15 can be expressed in a short form, like 10 < x < 15.
while and if keywords in the same program. Note the level of indentation
when the if statement comes inside the while loop. Remember that, the if statement must be
aligned with the corresponding elif and else.
The statement
Example:
big2.py
x = 1
while x < 11:
if x < 5:
print 'Small ', x
else:
print 'Big ', x
x = x + 1
print 'Done'
break
The
continue
statement is used to skip the rest of the block and go to the beginning again. Both are demonstrated
in the program
Example:
big3.py
shown below.
big3.py
x = 1
while x < 10:
if x < 3:
print 'skipping work', x
x = x + 1
continue
print x
if x == 4:
print 'Enough of work'
break
x = x + 1
print 'Done'
The output of big3.py is listed below.
skipping work 1
skipping work 2
3
4
Enough of work
Done
Now let us write a program to nd out the largest positive number entered by the user. The
algorithm works in the following manner.
zero. After reading a number, the program checks whether it is bigger than the current value of
the largest number. If so the value of the largest number is replaced with the current number. The
program terminates when the user enters zero.
also.
CHAPTER 2.
PROGRAMMING IN PYTHON
23
Example: max.py
max = 0
while True:
# Infinite loop
x = input('Enter a number ')
if x > max:
max = x
if x == 0:
print max
break
2.12 Exercises
We have now covered the minimum essentials of Python; dening variables, performing arithmetic
and logical operations on them and the control ow statements. These are sucient for handling
most of the programming tasks. It would be better to get a grip of it before proceeding further,
by writing some code.
1. Modify the expression
print 5+3*2
to get a result of 16
print type(4.5)
3. Print all even numbers upto 30, suxed by a * if the number is a multiple of 6. (hint: use
% operator)
4. Write Python code to remove the last two characters of 'I am a long string' by slicing, without
counting the characters. (hint: use negative indexing)
5. s = '012345' . (a) Slice it to remove last two elements (b) remove rst two element.
6. a = [1,2,3,4,5]. Use Slicing and multiplication to generate [2,3,4,2,3,4] from it.
CHAPTER 2.
24
PROGRAMMING IN PYTHON
cm3 (
1 inch =
2.54 cm)
14. Write a program to nd the percentage of volume occupied by a sphere of diameter
in a cube of side
r.
Read
tted
2 + 5j
and
2 5j
dening the real and imaginary parts separately and using the multiplication formula.
CHAPTER 2.
25
PROGRAMMING IN PYTHON
2.13 Functions
Large programs need to be divided into small logical units.
unit of code that has a name and does a well dened job. A function groups a number of program
statements into a unit and gives it a name. This unit can be invoked from other parts of a program.
Python allows you to dene functions using the
def
variables as arguments, which receive their values from the calling program.
In the example shown below, function arguments (a and b) get the values 3 and 4 respectively
from the caller. One can specify more than one variables in the return statement, separated by
commas.
have any arguments, but while calling them we need to use an empty parenthesis, otherwise the
function will not be invoked. If there is no return statement, a None is returned to the caller.
Example func.py
def sum(a,b):
# a trivial function
return a + b
print sum(3, 4)
The function
factorial.py
call. Try to understand the working of this by inserting print statements inside the function.
Example factor.py
CHAPTER 2.
26
PROGRAMMING IN PYTHON
Example banocci.py
def fib(n): # print Fibonacci series up to n
a, b = 0, 1
while b < n:
print b
a, b = b, a+b
print fib(30)
Runing
banocci.py
will print
1 1 2 3 5 8 13 21
Modify it to replace
a, b = b, a + b
third variable.
scope.py
demonstrates
this feature.
Example scope.py
def change(x):
counter = x
counter = 10
change(5)
print counter
The program will print 10 and not 5.
to each other. In some cases, it may be desirable to allow the function to change some external
variable. This can be achieved by using the
global
keyword, as shown in
global.py.
Example global.py
def change(x):
global counter # use the global variable
counter = x
counter = 10
change(5)
print counter
The program will now print 5. Functions with global variables should be used carefully to avoid
inadvertent side eects.
CHAPTER 2.
27
PROGRAMMING IN PYTHON
Example power.py
def power(mant, exp = 2.0):
return mant ** exp
print power(5., 3)
print power(4.)
print power()
# prints 16
# Gives Error
Example named.py
def power(mant = 10.0, exp = 2.0):
return mant ** exp
print power(5., 3)
print power(4.)
print power(exp=3)
# prints 16
# mant gets 10.0, prints 1000
Example: stringhelp.py
s = 'hello world'
print len(s)
print s.upper()
help(str)
# press q to exit help
Python is an object oriented language and all variables are objects belonging to various classes.
The method upper() (a function belonging to a class is called a method) is invoked using the dot
operator. All we need to know at this stage is that there are several methods that can be used for
manipulating objects and they can be invoked like:
variable_name.method_name().
CHAPTER 2.
28
PROGRAMMING IN PYTHON
Example: split.py
s = 'I am a long string'
print s.split()
a = 'abc.abc.abc'
aa = a.split('.')
print aa
mm = '+'.join(aa)
print mm
The result is shown below
['I', 'am', 'a', 'long', 'string']
['abc', 'abc', 'abc']
'abc+abc+abc'
The List of strings generated by split is joined using '+' character, resulting in the last line of the
output.
list3.py
Example: list3.py
a = []
# make an empty list
a.append(3)
# Add an element
a.insert(0,2.5) # insert 2.5 as first element
print a, a[0]
print len(a)
The output is shown below.
[2.5, 3] 2.5
2
b=a
from list a, it just make a reference to a. The following example will clarify this point. To make a
duplicate copy of a list, we need to use the
Example: list_copy.py
copy
module.
CHAPTER 2.
29
PROGRAMMING IN PYTHON
a = [1,2,3,4]
print a
b = a
# b refers to a
print a == b # True
b[0] = 5
# Modifies a[0]
print a
import copy
c = copy.copy(a)
c[1] = 100
print a is c # is False
print a, c
The output is shown below.
[1, 2, 3, 4]
True
[5, 2, 3, 4]
False
[5, 2, 3, 4] [5, 100, 3, 4]
One of the major advantages of Python is the availability of libraries for various applications like
graphics, networking and scientic computation.
itself has a large number of modules: time, random, pickle, system etc. are some of them. The
site http://docs.python.org/library/ has the complete reference.
Modules are loaded by using the
import
import
is explained
module_name.function_name().
Example mathsin.py
import math
print math.sin(0.5) # module_name.method_name
Example mathsin2.py
import math as m
print m.sin(0.5)
We can also import the functions to behave like local (like the ones within our source le) function,
as shown below. The character * is a wild card for importing all the functions.
5 While giving names to your Python programs, make sure that you are not directly or indirectly importing any
Python module having same name. For example, if you create a program named math.py and keep it in your working
directory, the import math statement from any other program started from that directory will try to import your
le named math.py and give error. If you ever do that by mistake, delete all the les with .pyc extension from your
directory.
CHAPTER 2.
30
PROGRAMMING IN PYTHON
Example mathlocal.py
from math import sin # sin is imported as local
print sin(0.5)
Example mathlocal2.py
from math import *
print sin(0.5)
In the third and fourth cases, we need not type the module name every time. But there could be
conict.py,
numpy is capable of handling a list argument. After importing math.py, line 4, the
sin function from math module replaces the one from numpy. The error occurs because the sin()
from math can accept only a numeric type argument.
trouble if two modules imported contains a function with same name. In the program
the
sin()
from
Example conict.py
from numpy import *
x = [0.1, 0.2, 0.3]
print sin(x)
from math import *
print sin(x)
2.15.2 Packages
Packages are used for organizing multiple modules. The module name A.B designates a submodule
6 and
named B in a package named A. The concept is demonstrated using the packages Numpy
Scipy.
Example submodule.py
import numpy
print numpy.random.normal()
import scipy.special
print scipy.special.j0(.1)
random is a module inside the package NumPy. Similarly special is a module inside
Scipy. We use both of them in the package.module.function() format. But there is
In this example
the package
some dierence. In the case of Numpy, the random module is loaded by default, importing scipy
does not import the module special by default. This behavior can be dened while writing the
Package and it is upto the package author.
Example wle.py
6 NumPy
CHAPTER 2.
PROGRAMMING IN PYTHON
31
f = open('test.dat' , 'w')
f.write ('This is a test file')
f.close()
Above program creates a new le named 'test.dat' (any existing le with the same name will be
deleted) and writes a String to it. The following program opens this le for reading the data.
Example rle.py
f = open('test.dat' , 'r')
print f.read()
f.close()
Note that the data written/read are character strings. read() function can also be used to read a
xed number of characters, as shown below.
Example rle2.py
f = open('test.dat' , 'r')
print f.read(7)
# get first seven characters
print f.read()
# get the remaining ones
f.close()
Now we will examine how to read a text data from a le and convert it into numeric type. First
we will create a le with a column of numbers.
Example wle2.py
f = open('data.dat' , 'w')
for k in range(1,4):
s = '%3d\n' %(k)
f.write(s)
f.close()
The contents of the le created will look like this.
1
2
3
Now we write a program to read this le, line by line, and convert the string type data to integer
Example rle3.py
f = open('data.dat' , 'r')
while 1: # infinite loop
s = f.readline()
if s == :
# Empty string means end of file
break
# terminate the loop
m = int(s)
# Convert to integer
print m * 5
f.close()
7 This will give error if there is a blank line in the data le. This can be corrected by changing the comparison
statement to if len(s) < 1: , so that the processing stops at a blank line. Modify the code to skip a blank line
instead of exiting (hint: use continue ).
CHAPTER 2.
PROGRAMMING IN PYTHON
32
Example pickledump.py
import pickle
f = open('test.pck' , 'w')
pickle.dump(12.3, f) # write a float type
f.close()
Now write another program to read it back from the le and check the data type.
Example pickleload.py
import pickle
f = open('test.pck' , 'r')
x = pickle.load(f)
print x , type(x)
# check the type of data read
f.close()
Example: format.py
a = 2.0 /3 # 2/3 will print zero
print a
print 'a = %5.3f' %(a) # upto 3 decimal places
Data can be printed in various formats.
table. There are several ags that can be used to modify the formatting, like justication, lling
etc.
The following example shows some of the features available with formatted printing.
Example: format2.py
a = 'justify as you like'
print '%30s'%a
# right justified
print '%-30s'%a
# minus sign for left justification
for k in range(1,11): # A good looking table
print '5 x %2d = %2d' %(k, k*5)
CHAPTER 2.
33
PROGRAMMING IN PYTHON
Conversion
Conversion
Example
Result
d , i
signed Integer
'%6d'%(12)
'%6.4f '%(2.0/3)
'
0.667
12'
'%6.2e'%(2.0/3)
6.67e-01
hexadecimal
'%x'%(16)
10
octal
'%o'%(8)
10
string
'%s'%('abcd')
abcd
0d
modied 'd'
'%05d'%(12)
00012
The output of
format2.py
is given below.
try
and
except
Example: except.py
x = input('Enter a number ')
try:
print 10.0/x
except:
print 'Division by zero not allowed'
If any exception occurs while running the code inside the try block, the code inside the except
block is executed. The following program implements error checking on input using exceptions.
Example: except2.py
def get_number():
while 1:
try:
a = raw_input('Enter a number ')
x = atof(a)
CHAPTER 2.
34
PROGRAMMING IN PYTHON
return x
except:
print 'Enter a valid number'
print get_number()
Pen
on a two di-
mensional screen to generate graphical patterns. The Pen can be controlled using functions like
Example turtle1.py
from turtle import *
a = Pen()
# Creates a turtle in a window
a.forward(50)
a.left(45)
a.backward(50)
a.right(45)
a.forward(50)
a.circle(10)
a.up()
a.forward(50)
a.down()
a.color('red')
a.right(90)
a.forward(50)
raw_input('Press Enter')
Example turtle2.py
from turtle import *
a = Pen()
for k in range(4):
a.forward(50)
a.left(90)
a.circle(25)
raw_input() # Wait for Key press
8 http://docs.python.org/library/turtle.html
CHAPTER 2.
35
PROGRAMMING IN PYTHON
Outputs of the program turtle2.py and turtle3.py are shown in gure 2.2.
Example turtle3.py
from turtle import *
def draw_rectangle():
for k in range(4):
a.forward(50)
a.left(90)
a = Pen()
for k in range(36):
draw_rectangle()
a.left(10)
raw_input()
The program turtle3.py creates a pattern by drwaing 36 squares, each drawn tilted by
10
from
the previous one. The program turtle4.py generates the fractal image as shown in gure2.2(c).
Example turtle4.py
from turtle import *
def f(length, depth):
if depth == 0:
forward(length)
else:
f(length/3, depth-1)
right(60)
f(length/3, depth-1)
left(120)
f(length/3, depth-1)
right(60)
f(length/3, depth-1)
f(500, 4)
raw_input('Press any Key')
CHAPTER 2.
36
PROGRAMMING IN PYTHON
of this chapter is just to show the ease of making GUI in Python and we have selected Tkinter ,
one of the easiest to learn. The GUI programs are event driven (movement of mouse, clicking a
mouse button, pressing and releasing a key on the keyboard etc. are called events). The execution
sequence of the program is decided by the events, generated mostly by the user. For example, when
the user clicks on a Button, the code associated with that Button is executed. GUI Programming
is about creating Widgets like Button, Label, Canvas etc. on the screen and executing selected
functions in response to events. After creating all the necessary widgets and displaying them on
the screen, the control is passed on to Tkinter by calling a function named
mainloop.
After that
the program ow is decided by the events and associated callback functions.
For writing GUI programs, the rst step is to create a main graphics window by calling the
function Tk(). After that we create various Widgets and pack them inside the main window. The
example programs given below demonstrate the usage of some of the Tkinter widgets.The program
tkmain.py
is the smallest GUI program one can write using Tkinter. The output of tkmain.py is
shown in gure2.3(a).
Example tkmain.py
from Tkinter import *
root = Tk()
root.mainloop()
Example tklabel.py
from Tkinter import *
root = Tk()
w = Label(root, text="Hello, world")
w.pack()
root.mainloop()
The program tklabel.py will generate the output as shown in gure 2.3(b). Terminate the program
by clicking on the x displayed at the top right corner. In this example, we used a Label widget to
display some text. The next example will show how to use a Button widget.
9 http://www.pythonware.com/library/an-introduction-to-tkinter.htm
http://infohost.nmt.edu/tcc/help/pubs/tkinter/
http://wiki.python.org/moin/TkInter
CHAPTER 2.
PROGRAMMING IN PYTHON
37
A Button widget can have a callback function, hello() in this case, that gets executed when the
user clicks on the Button. The program will display a Button on the screen. Every time you click
on it, the function
hello
Example tkbutton.py
from Tkinter import *
def hello():
print 'hello world'
w = Tk() # Creates the main Graphics window
b = Button(w, text = 'Click Me', command = hello)
b.pack()
w.mainloop()
Canvas is another commonly used widget. Canvas is a drawing area on which we can draw elements
like line, arc, rectangle, text etc. The program tkcanvas.py creates a Canvas widget and binds the
<Button-1> event to the function draw(). When left mouse button is pressed, a small rectangle
are drawn at the cursor position. The output of the program is shown in gure 2.4(b).
Example tkcanvas.py
from Tkinter import *
def draw(event):
c.create_rectangle(event.x, \
event.y, event.x+5, event.y+5)
w = Tk()
c = Canvas(w, width = 300, height = 200)
c.pack()
c.bind("<Button-1>", draw)
w.mainloop()
The next program is a modication of tkcanvas.py. The right mouse-button is bound to remove().
Every time a rectangle is drawn, its return value is added to a list, a global variable, and this list
is used for removing the rectangles when right button is pressed.
CHAPTER 2.
PROGRAMMING IN PYTHON
38
Example tkcanvas2.py
from Tkinter import *
recs = []
# List keeping track of the rectangles
def remove(event):
global recs
if len(recs) > 0:
c.delete(recs[0]) # delete from Canvas
recs.pop(0)
# delete first item from list
def draw(event):
global recs
r = c.create_rectangle(event.x, \
event.y, event.x+5, event.y+5)
recs.append(r)
w = Tk()
c = Canvas(w, width = 300, height = 200)
c.pack()
c.bind("<Button-1>", draw)
c.bind("<Button-3>", remove)
w.mainloop()
and their interactions to design computer programs. Python is an object oriented language but it
does not force you to make all programs object oriented and there is no advantage in making small
programs object oriented. In this section, we will discuss some features of OOP.
Before going to the new concepts, let us recollect some of the things we have learned. We have
seen that the eect of operators on dierent data types is predened. For example 2 3 results in
6 and 2 0 abc0 results in 0 abcabc0 . This behavior has been decided beforehand, based on some logic,
by the language designers. One of the key features of OOP is the ability to create user dened
data types. The user will specify, how the new data type will behave under the existing operators
like add, subtract etc. and also dene methods that will belong to the new data type.
We will design a new data type using the class keyword and dene the behavior of it. In the
program point.py, we dene a class named Point. The variables xpos and ypos are members of
Point.
The __init__() function is executed whenever we create an instance of this class, the
member variables are initialized by this function. The way in which an object belonging to this
class is printed is decided by the __str__ function. We also have dened the behavior of add (+)
and subtract (-) operators for this class. The + operator returns a new Point by adding the x and
y coordinates of two Points. Subtracting a Point from another gives the distance between the two.
The method dist() returns the distance of a Point object from the origin. We have not dened the
behavior of Point under copy operation. We can use the copy module of Python to copy objects.
Example point.py
class Point:
'
CHAPTER 2.
PROGRAMMING IN PYTHON
39
Example point1.py
from point import *
origin = Point()
print origin
p1 = Point(4,4)
p2 = Point(8,7)
print p1
print p2
print p1 + p2
print p1 - p2
print p1.dist()
Output of program point1.py is shown below.
Point at (0.000000,0.000000)
Point at (4.000000,4.000000)
Point at (8.000000,7.000000)
Point at (12.000000,11.000000)
5.0
5.65685424949
In this section, we have demonstrated the OO concepts like class, object and operator overloading.
CHAPTER 2.
40
PROGRAMMING IN PYTHON
Example cpoint.py
class colPoint(Point): #colPoint inherits Point
color = 'black'
def __init__(self,x=0,y=0,col='black'):
Point.__init__(self,x,y)
self.color = col
def __str__(self):
return '%s colored Point at (%f,%f)'% \
(self.color,self.xpos, self.ypos)
Example point2.py
from cpoint import *
p1 = Point(5,5)
rp1 = colPoint(2,2,'red')
print p1
print rp1
print rp1 + p1
print rp1.dist()
The output of point2.py is listed below.
Point at (5.000000,5.000000)
red colored Point at (2.000000,2.000000)
Point at (7.000000,7.000000)
2.82842712475
For a detailed explanation on the object oriented features of Python, refer to chapters 13, 14 and
15 of the online book http://openbookproject.net//thinkCSpy/
conversion is performed internally. The method named line() accepts a list of xy coordinates. The
le
tkplot_class.py
Example tkplot_class.py
from Tkinter import *
from math import *
class disp:
def __init__(self, parent, width=400., height=200.):
CHAPTER 2.
41
PROGRAMMING IN PYTHON
self.parent = parent
self.SCX = width
self.SCY = height
self.border = 1
self.canvas = Canvas(parent, width=width, height=height)
self.canvas.pack(side = LEFT)
self.setWorld(0 , 0, self.SCX, self.SCY) # scale factors
def setWorld(self, x1, y1, x2, y2):
self.xmin = float(x1)
self.ymin = float(y1)
self.xmax = float(x2)
self.ymax = float(y2)
self.xscale = (self.xmax - self.xmin) / (self.SCX)
self.yscale = (self.ymax - self.ymin) / (self.SCY)
def w2s(self, p): #world-to-screen before plotting
ip = []
for xy in p:
ix = self.border + int( (xy[0] - self.xmin) / self.xscale)
iy = self.border + int( (xy[1] - self.ymin) / self.yscale)
iy = self.SCY - iy
ip.append((ix,iy))
return ip
def line(self, points, col='blue'):
ip = self.w2s(points)
t = self.canvas.create_line(ip, fill=col)
The program
tkplot.py
imports tkplot_class.py and plots two graphs. The advantage of code reuse
10 . Output of
tkplot.py
Example tkplot.py
from tkplot_class import *
from math import *
w = Tk()
gw1 = disp(w)
xy = []
10 A more sophisticated version of the
available on the CD.
disp
CHAPTER 2.
PROGRAMMING IN PYTHON
42
for k in range(200):
x = 2 * pi * k/200
y = sin(x)
xy.append((x,y))
gw1.setWorld(0, -1.0, 2*pi, 1.0)
gw1.line(xy)
gw2 = disp(w)
gw2.line([(10,10),(100,100),(350,50)], 'red')
w.mainloop()
2.22 Exercises
1. Generate multiplication table of eight and write it to a le.
2. Make a list and write it to a le using the pickle module.
3. Write a Python program to open a le and write 'hello world' to it.
4. Write a Python program to open a text le and read all lines from it.
5. Write a program to generate the multiplication table of a number from the user. The output
should be formatted as shown below
1 x 5 =
2 x 5 = 10
6. Dene the list [1,2,3,4,5,6] using the range function. Write code to insert a 10 after 2, delete
4, add 0 at the end and sort it in the ascending order.
7. Write Python code to generate the sequence of numbers
25 20 15 10 5
using range function . Delete 15 from the result and sort it. Print it using a for loop.
8. Dene a string
CHAPTER 2.
43
PROGRAMMING IN PYTHON
15. Write a program to nd the length of the hypotenuse of a right triangle from the length of
other two sides, get the input from the user.
16. Write a program displaying 2 labels and 2 buttons. It should print two dierent messages
when clicked on the two buttons.
17. Write a program with a Canvas and a circle drawn on it.
18. Write a program using for loop to reverse a string.
19. Write a Python function to calculate the GCD of two numbers
20. Write a program to print the values of sine function from
to
x = (b
b2 4ac)/2a
3x2 + 6x + 12 = 0
23. Write a program to calculate the distance between points (x1,y1) and (x2,y2) in a Cartesian
plane. Get the coordinates from the user.
24. Write a program to evaluate
y=
2.3a + a2 + 34.5
for a = 1, 2 and 3.
25. Print Fibanocci numbers upto 100, without using multiple assignment statement.
26. Draw a chess board pattern using turtle graphics.
27. Find the syntax error in the following code and correct it.
x=1
while x <= 10:
print x * 5
Chapter 3
math
module to calculate trigonometric functions. Using the tools introduced so far, let us
generate the data points to plot a sine wave. The program sine.py generates the coordinates to
plot a sine wave.
Example sine.py
import math
x = 0.0
while x < 2 * math.pi:
print x , math.sin(x)
x = x + 0.1
The output to the screen can be redirected to a le as shown below, from the command prompt.
You can plot the data using some program like xmgrace.
$ python sine.py > sine.dat
$ xmgrace sine.dat
It would be better if we could write such programs without using loops explicitly.
scientic computing requires manipulating of large data structures like matrices.
The
Serious
list
data
type of Python is very exible but the performance is not acceptable for large scale computing.
The need of special tools is evident even from the simple example shown above.
NumPy
is a
for creating matrices. The mathematical functions like sine, cosine etc. of numpy accepts array
objects as arguments and return the results as arrays objects.
sliced and copied like Python Lists.
1 http://numpy.scipy.org/
http://www.scipy.org/Tentative_NumPy_Tutorial
http://www.scipy.org/Numpy_Functions_by_Category
http://www.scipy.org/Numpy_Example_List_With_Doc
44
CHAPTER 3.
In the examples below, we will import numpy functions as local (using the syntax
import * ).
45
from numpy
Since it is the only package used there is no possibility of any function name conicts.
Example numpy1.py
from numpy import *
x = array( [1, 2, 3] )
print x , type(x)
Example numpy3.py
from numpy import *
a = [ [1,2] , [3,4] ] # make a list of lists
x = array(a)
# and convert to an array
print a
Other than than
array(), there are several other functions that can be used for creating dierent
0.0
0.0
0.0 0.0
0.0 0.0
CHAPTER 3.
46
3.1.1.5 random.random(shape)
Similar to the functions above, but the matrix is lled with random numbers ranging from 0 to 1,
of
oat
reshape()
Working of
reshape()
reshape.py
Example reshape.py
from numpy import *
a = arange(20)
b = reshape(a, [4,5])
print b
The result is shown below.
array([[ 0,
[ 5,
[10,
[15,
The program
1,
6,
11,
16,
numpy2.py
2,
7,
12,
17,
3,
8,
13,
18,
The function
4],
9],
14],
19]])
Example numpy2.py
from numpy import *
a = arange(1.0, 2.0, 0.1)
print a
b = linspace(1,2,11)
print b
c = ones(5,'float')
print c
d = zeros(5, 'int')
print d
e = random.rand(5)
print e
Output of this program will look like;
[ 1. 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9]
[ 1. 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2. ]
[ 1. 1. 1. 1. 1.]
[ 0. 0. 0. 0. 0.]
[ 0.89039193 0.55640332 0.38962117 0.17238343 0.01297415]
CHAPTER 3.
47
3.1.2 Copying
Numpy arrays can be copied using the copy method, as shown below.
Example array_copy.py
from mumpy import *
a = zeros(5)
print a
b = a
c = a.copy()
c[0] = 10
print a, c
b[0] = 10
print a,c
The output of the program is shown below.
The statement
b = a
aroper.py
Example aroper.py
from numpy import *
a = array([[2,3], [4,5]])
b = array([[1,2], [3,0]])
print a + b
print a * b
The output will be as shown below
array([[3, 5],
[7, 5]])
array([[ 2, 6],
[12, 0]])
Modifying this program for more operations is left as an exercise to the reader.
i
A B = A1
B1
j
A2
B2
k
A3
B3
= i(A2 B3 A3 B2 ) + j(A1 B3 A3 B1 ) + k(A1 B2 A2 B1 )
(3.1)
CHAPTER 3.
48
cross.py
prints [-3,
6, -3]
Example cross.py
from numpy import *
a = array([1,2,3])
b = array([4,5,6])
c = cross(a,b)
print c
cross.py
to
c = dot(a, b),
A.B = A1 B1 + A2 B2 + A3 B3
. If you change
array=fromle()
array.tole(lename)
Example leio.py
from numpy import *
a = arange(10)
a.tofile('myfile.dat')
b = fromfile('myfile.dat',dtype = 'int')
print b
The function fromle() sets dtype='oat' by default. In this case we have saved an integer array
and need to specify that while reading the le. We could have saved it as oat the the statement
a.tole('myle.dat', 'oat').
We can verify
argument should normally result in an error message. In some cases, you may get a result whose
elements are having very high values, and it indicates an error.
Example inv.py
from numpy import *
a = array([ [4,1,-2], [2,-3,3], [-6,-2,1] ], dtype='float')
ainv = linalg.inv(a)
print ainv
print dot(a,ainv)
Result of this program is printed below.
CHAPTER 3.
49
Example vfunc.py
from numpy import *
a = array([1,10,100,1000])
print log10(a)
The output of the program is [ 0.
1.
2.
3.]
returned in an array. This feature simplies the programs a lot. Numpy also provides a function
to vectorize functions written by the user.
Example vectorize.py
from numpy import *
def spf(x):
return 3*x
vspf = vectorize(spf)
a = array([1,2,3,4])
print vspf(a)
The output will be [ 3 6 9 12] .
3.3 Exercises
1. Write code to make a one dimensional matrix with elements 5,10,15,20 and 25. make another
matrix by slicing the rst three elements from it.
2. Create a
32
matrix and print the sum of its elements using for loops.
3. Create a
23
b.
CHAPTER 3.
50
= 2x2 + x3 = 11
=2x1 + 4x2 = 2x3 = =16
x1 = 2x2 + 4x3 = 17
4x1
11. Find the new values of the coordinate (10,10) under a rotation by angle
12. Write a vectorized function to evaluate
y = x2
/4.
Chapter 4
Data visualization
A graph or chart is used to present numerical data in visual form. A graph is one of the easiest
ways to compare numbers. They should be used to make facts clearer and more understandable.
Results of mathematical computations are often presented in graphical format. In this chapter, we
will explore the Python modules used for generating two and three dimensional graphs of various
types.
pylab, the plotting functions from the submodules pyplot and matrix
mlab will be available as local functions. Pylab also
imports Numpy for you. Let us start with some simple plots to become familiar with matplotlib.
Example plot1.py
from pylab import *
data = [1,2,5]
plot(data)
show()
In the above example, the x-axis of the three points is taken from 0 to 2. We can specify both the
axes as shown below.
Example plot2.py
from pylab import *
x = [1,2,5]
y = [4,5,6]
1 http://matplotlib.sourceforge.net/
http://matplotlib.sourceforge.net/users/pyplot_tutorial.html
http://matplotlib.sourceforge.net/examples/index.html
http://matplotlib.sourceforge.net/api/axes_api.html
51
CHAPTER 4.
DATA VISUALIZATION
52
plot(x,y)
show()
By default, the color is blue and the line style is continuous. This can be changed by an optional
argument after the coordinate data, which is the format string that indicates the color and line
type of the plot. The default format string is `b-` (blue, continuous line). Let us rewrite the above
example to plot using red circles. We will also set the ranges for x and y axes and label them.
Example plot3.py
from pylab import *
x = [1,2,5]
y = [4,5,6]
plot(x,y,'ro')
xlabel('x-axis')
ylabel('y-axis')
axis([0,6,1,7])
show()
The gure 4.1 shows two dierent plots in the same window, using dierent markers and colors.
Example plot4.py
from pylab import *
t = arange(0.0, 5.0, 0.2)
plot(t, t**2,'x')
# t2
plot(t, t**3,'ro')
# t3
show()
We have just learned how to draw a simple plot using the pylab interface of matplotlib.
Example subplot1.py
CHAPTER 4.
53
DATA VISUALIZATION
The arguments to subplot function are NR (number of rows) , NC (number of columns) and a
gure number, that ranges from 1 to
if
N R N C < 10,
N R N C.
subplot2.py.
to watch the results. Please note that the % character has dierent meanings. In
(pn+1)%5,
it
is the reminder operator resulting in a number less than 5. The % character also appears in the
String formatting.
Example subplot2.py
from pylab import *
mark = ['x','o','^','+','>']
NR = 2 # number of rows
NC = 3 # number of columns
pn = 1
for row in range(NR):
for col in range(NC):
subplot(NR, NC, pn)
a = rand(10) * pn
plot(a, marker = mark[(pn+1)%5])
xlabel('plot %d X'%pn)
ylabel('plot %d Y'%pn)
pn = pn + 1
show()
The angle
x axis
Use the
x axis
Matplotlib supports polar plots, using the polar(, r ) function. Let us plot a circle using polar().
Example polar.py
from pylab import *
th = linspace(0,2*pi,100)
r = 5 * ones(100) # radius = 5
polar(th,r)
show()
CHAPTER 4.
54
DATA VISUALIZATION
Example piechart.py
from pylab import *
labels = 'Frogs', 'Hogs', 'Dogs', 'Logs'
fracs = [25, 25, 30, 20]
pie(fracs, labels=labels)
show()
arange, linspace
and
logspace
functions from
numpy
to generate the
input data and also the vectorized versions of the mathematical functions. For arange(), the third
argument is the stepsize. The total number of elements is calculated from start, stop and stepsize.
In the case of linspace(), we provide start, stop and the total number of points. The step size is
calculated from these three parameters. Please note that to create a data set ranging from 0 to 1
(including both) with a stepsize of 0.1, we need to specify linspace(0,1,11) and not linspace(0,1,10).
Let the rst example be the familiar sine function. The input data is from to + radians . To
2
make it a bit more interesting we are plotting sin x also. The objective is to explain the concept
of odd and even functions. Mathematically, we say that a function
and is odd if
f (x) = f (x).
f (x)
is even if
f (x) = f (x)
Even functions are functions for which the left half of the plane
looks like the mirror image of the right half of the plane. From the gure 4.2(a) you can see that
sin x is odd and sin x2 is even.
Example npsin.py
2 Why do we need to give the angles in radians and not in degrees. Angle in radian is the length of the arc dened
by the given angle, with unit radius. Degree is just an arbitrary unit.
CHAPTER 4.
55
DATA VISUALIZATION
npsin.py
to plot
etc.
x2 + y 2 = a2
of the coordinate system. In order to plot it using Cartesian coordinates, we need to express
terms of
x,
and is given by
y=
in
a2 x2
of y. This will give us only half of the circle, since for each value of x, there are two values of y
(+y and -y). The following program
circ.py
gure 4.2(b). Any multi-valued function will have this problem while plotting. Such functions can
be plotted better using parametric equations or using the polar plot options, as explained in the
coming sections.
Example circ.py
from pylab import *
a = 10.0
x = linspace(-a, a , 200)
yupper = sqrt(a**2 - x**2)
ylower = -sqrt(a**2 - x**2)
plot(x,yupper)
plot(x,ylower)
show()
x = a cos
and
y = a sin
Example circpar.py
from pylab import *
a = 10.0
th = linspace(0, 2*pi, 200)
x = a * cos(th)
y = a * sin(th)
plot(x,y)
show()
CHAPTER 4.
56
DATA VISUALIZATION
to less than
, 1.5and2
The
for
also depends on the loop variable. For the next three values it will
Example arcs.py
from pylab import *
a = 10.0
for a in range(5,21,5):
th = linspace(0, pi * a/10, 200)
x = a * cos(th)
y = a * sin(th)
plot(x,y)
show()
A collection of them is
available on the Internet [2][3]. We will select some of them and plot here. Exploring them further
is left as an exercise to the reader.
4.3.1 Astroid
The astroid was rst discussed by Johann Bernoulli in 1691-92. It also appears in Leibniz's correspondence of 1715. It is sometimes called the tetracuspid for the obvious reason that it has four
cusps. A circle of radius 1/4 rolls around inside a circle of radius 1 and a point on its circumference
traces an astroid. The Cartesian equation is
2
x3 + y 3 = a3
(4.1)
(4.2)
CHAPTER 4.
57
DATA VISUALIZATION
In order to plot the curve in the Cartesian system, we rewrite equation 4.1 as
2
y = (a 3 x 3 ) 2
The program
The program
astropar.py
uses the parametric equation and plots the complete curve. Both are shown in gure 4.4
Example astro.py
from pylab import *
a = 2
x = linspace(0,a,100)
y = ( a**(2.0/3) - x**(2.0/3) )**(3.0/2)
plot(x,y)
show()
Example astropar.py
from pylab import *
a = 2
t = linspace(-2*a,2*a,101)
x = a * cos(t)**3
y = a * sin(t)**3
plot(x,y)
show()
4.3.2 Ellipse
The ellipse was rst studied by Menaechmus [4] . Euclid wrote about the ellipse and it was given
its present name by Apollonius. The focus and directrix of an ellipse were considered by Pappus.
Kepler, in 1602, said he believed that the orbit of Mars was oval, then he later discovered that it
was an ellipse with the sun at one focus. In fact Kepler introduced the word
focus
and published
x2
y2
+
=1
a2
b2
(4.3)
x = a cos(t), y = b sin(t)
(4.4)
The program
ellipse.py
uses the parametric equation to plot the curve. Modifying the para-
metric equations will result in Lissajous gures. The output of lissa.py are shown in gure 4.4(c).
CHAPTER 4.
58
DATA VISUALIZATION
Example ellipse.py
from pylab import *
a = 2
b = 3
t = linspace(0, 2 * pi, 100)
x = a * sin(t)
y = b * cos(t)
plot(x,y)
show()
Example lissa.py
from pylab import *
a = 2
b = 3
t= linspace(0, 2*pi,100)
x = a * sin(2*t)
y = b * cos(t)
plot(x,y)
x = a * sin(3*t)
y = b * cos(2*t)
plot(x,y)
show()
The Lissajous curves are closed if the ratio of the arguments for sine and cosine functions is an
integer. Otherwise open curves will result, both are shown in gure 4.4(c).
r2 = a2 .
r = a.
Example archi.py
from pylab import *
a = 2
CHAPTER 4.
59
DATA VISUALIZATION
Example fermat.py
from pylab import *
a = 2
th= linspace(0, 10*pi,200)
r = sqrt(a**2 * th)
polar(th,r)
polar(th, -r)
show()
2k
petals and
an even
curve is
closed and has nite length. If k is irrational, then it is not closed and has innite length.
Example rose.py
from pylab import *
k = 4
th = linspace(0, 10*pi,1000)
r = cos(k*th)
polar(th,r)
show()
There are dozens of other famous curves whose details are available on the Internet. It may be an
interesting exercise for the reader. For more details refer to [3, 2, 5]on the Internet.
d2 y
+y =0
dx2
which has a series solution of the form
y = a0
(1)
n=0
2n+1
x2n
n x
+ a1
(1)
(2n)!
(2n + 1)!
n=0
(4.5)
These are the Maclaurin series for sine and cosine functions. The following code plots several
terms of the sine series and their sum.
CHAPTER 4.
60
DATA VISUALIZATION
Example series_sin.py
from pylab import *
from scipy import factorial
x = linspace(-pi, pi, 50)
y = zeros(50)
for n in range(5):
term = (-1)**(n) * (x**(2*n+1)) / factorial(2*n+1)
y = y + term
#plot(x,term) #uncomment to see each term
plot(x, y, '+b')
plot(x, sin(x),'r') # compare with the real one
show()
The output of
series_sin.py
library is plotted.
sin
The values calculated by using the series becomes closer to the actual value
with more and more number of terms. The error can be obtained by adding the following lines to
series_sin.py
err = y - sin(x)
plot(x,err)
for k in err:
print k
f (x)
and cosines. The computation and study of Fourier series is known as harmonic analysis and is
extremely useful as a way to break up an arbitrary periodic function into a set of simple terms that
can be plugged in, solved individually, and then recombined to obtain the solution to the original
problem or an approximation to it to whatever accuracy is desired or practical.
The examples below shows how to generate a square wave and sawtooth wave using this technique. To make the output better, increase the number of terms by changing the argument of the
range() function, used in the for loop. The output of the programs are shown in gure 4.7.
Example fourier_square.py
CHAPTER 4.
61
DATA VISUALIZATION
Figure 4.7: Sawtooth and Square waveforms generated using Fourier series.
Example fourier_sawtooth.py
from pylab import *
N = 100
# number of points
x = linspace(-pi, pi, N)
y = zeros(N)
for n in range(1,10):
term = (-1)**(n+1) * sin(n*x) / n
y = y + term
plot(x,y)
show()
Example imshow1.py
from pylab import *
m = random([50,50])
imshow(m)
show()
imshow()
50 50
matrix lled
CHAPTER 4.
62
DATA VISUALIZATION
4.7 Fractals
3 are a part of fractal geometry, which is a branch of mathematics concerned with irregular
Fractals
patterns made of parts that are in some way similar to the whole (e.g.: twigs and tree branches).
A fractal is a design of innite details. It is created using a mathematical formula. No matter how
closely you look at a fractal, it never loses its detail. It is innitely detailed, yet it can be contained
in a nite space. Fractals are generally self-similar and independent of scale. The theory of fractals
was developed from Benoit Mandelbrot's study of complexity and chaos. Complex numbers are
required to compute the Mandelbrot and Julia Set fractals and it is assumed that the reader is
familiar with the basics of complex numbers.
To compute the basic Mandelbrot (or Julia) set one uses the equation
f (z) z 2 + c
, where
both z and c are complex numbers. The function is evaluated in an iterative manner, ie. the result
is assigned to
behavior of the values that are put into the function. If the value of the function goes to innity
(practically to some xed value, like 1 or 2) after few iterations for a particular value of
, that
point is considered to be outside the Set. A Julia set can be dened as the set of all the complex
2
numbers (z) such that the iteration of f (z) z + c is bounded for a particular value of c.
To generate the fractal the number of iterations required to diverge is calculated for a set of
points in the selected region in the complex plane. The number of iterations taken for diverging
decides the color of each point. The points that did not diverge, belonging to the set, are plotted
with the same color.
The program
julia.py
The program
creates a 2D array (200 x 200 elements). For our calculations, this array represents a rectangular
region on the complex plane centered at the origin whose lower left corner is (-1,-j) and the upper
right corner is (1+j). For 200x200 equidistant points in this plane the number of iterations are
calculated and that value is given to the corresponding element of the 2D matrix. The plotting is
taken care by the imshow function. The output is shown in gure 4.8(b). Change the value of
and run the program to generate more patterns. The equation also may be changed.
Example julia.py
'
Region of a complex plane ranging
and imaginary axes is represented
having X x Y elements.For X and Y
in the complex plane is 2.0/200 =
The nature of the pattern depends
'
from pylab import *
3 http://en.wikipedia.org/wiki/Fractal
CHAPTER 4.
63
DATA VISUALIZATION
X = 200
Y = 200
MAXIT = 100
MAXABS = 2.0
c = 0.02 - 0.8j # The constant in equation z**2 + c
m = zeros([X,Y]) # A two dimensional array
def numit(x,y):
# number of iterations to diverge
z = complex(x,y)
for k in range(MAXIT):
if abs(z) <= MAXABS:
z = z**2 + c
else:
return k
# diverged after k trials
return MAXIT
# did not diverge,
for x in range(X):
for y in range(Y):
re = 0.01 * x - 1.0
im = 0.01 * y - 1.0
m[x][y] = numit(re,im)
imshow(m) # Colored plot using
show()
4.8 Meshgrids
in order to make contour and 3D plots, we need to understand the meshgrid. Consider a rectangular
area on the X-Y plane. Assume there are m divisions in the X direction and n divisions in the Y
direction. We now have a
mn
coordinates of each mesh point is held in one matrix and y coordinates are held in another.
The NumPy function meshgrid() creates two 2x2 matrices from two 1D arrays, as shown in the
example below. This can be used for plotting surfaces and contours, by assigning a Z coordinate
to every mesh point.
Example mgrid1.py
from numpy import *
x = arange(0, 3, 1)
y = arange(0, 3, 1)
gx, gy = meshgrid(x, y)
print gx
print gy
The outputs are as shown below, gx(i,j) contains the x-coordinate and gx(i,j) contains the ycoordinate of the point (i,j).
[[0 1 2]
[0 1 2]
[0 1 2]]
[[0 0 0]
[1 1 1]
CHAPTER 4.
DATA VISUALIZATION
64
[2 2 2]]
We can evaluate a function at all points of the meshgrid by passing the meshgrid as an argument.
The program mgrid2.py plots the sum of sines of the x and y coordinates, using imshow to get a
result as shown in gure 4.8(c).
Example mgrid2.py
from pylab import *
x = arange(-3*pi, 3*pi, 0.1)
y = arange(-3*pi, 3*pi, 0.1)
xx, yy = meshgrid(x, y)
z = sin(xx) + sin(yy)
imshow(z)
show()
4.9 3D Plots
Matplotlib supports several types of 3D plots, using the Axes3D class. The following three lines
of code are required in every program making 3D plots using matplotlib.
Example sufrace3d.py
from pylab import *
from mpl_toolkits.mplot3d import Axes3D
ax = Axes3D(figure())
x = arange(-3*pi, 3*pi, 0.1)
y = arange(-3*pi, 3*pi, 0.1)
xx, yy = meshgrid(x, y)
z = sin(xx) + sin(yy)
ax.plot_surface(xx, yy, z, cmap=cm.jet, cstride=1)
show()
Example line3d.py
from pylab import *
from mpl_toolkits.mplot3d import Axes3D
ax = Axes3D(figure())
CHAPTER 4.
DATA VISUALIZATION
65
1
5
0
2
using mayavi. The plot of Y2 =
4
(3 cos 1) is shown in gure4.9.
Example ylm20.py
CHAPTER 4.
66
DATA VISUALIZATION
4.11 Exercises
1. Plot a sine wave using markers +, o and x using three dierent colors.
2. Plot
tan
from
from
to
2 ,
2
2ex /2 Frequency curve
b) y =
c)
d)
a cosh(x/a) catenary
sin(a) for a = 2, 3, and
4. Rhodonea curves
y=
n=
n=1
(1)n x2n+1
for 10 terms.
(2n+1)!
7. Write a Python program to calculate sine function using series expansion and plot it.
8. Write a Python program to plot
y = 5x2 + 3x + 2
with axes and title. Use red colored circles to mark the points.
CHAPTER 4.
67
DATA VISUALIZATION
9. Write a Python program to plot a Square wave using Fourier series, number of terms should
be a variable.
10. Write a Python program to read the x and y coordinates from a le, in a two column format,
and plot them.
11. Plot
x2 + y 2 + z 2 = 25
using mayavi.
y = x2 ,
4to4
Chapter 5
You prepare your document using a plain text editor, and the formatting is
specied by commands embedded in your document. The appearance of your document is decided
AT X, but you need to specify it using some commands. In this chapter, we will discuss some
by L
E
1
of these commands mainly to typeset mathematical equations.
hello.tex
Example hello.tex
\documentclass{article}
\begin{document}
Small is beautiful. % I am just a comment
\end{document}
Compile, view and make a PDF le using the following commands:
$ latex hello.tex
$ xdvi hello.dvi
$ dvipdf hello.dvi
The output will look like : Small is beautiful.
1 http://www.latex-project.org/
http://mirror.ctan.org/info/lshort/english/lshort.pdf
http://en.wikibooks.org/wiki/LATEX
68
CHAPTER 5.
69
texts.tex
Example texts.tex
\documentclass{article}
\begin{document}
This is normal text.
\newline
\textbf{This is bold face text.}
\textit{This is italic text.}\\
\tiny{This is tiny text.}
\large{This is large text.}
\underline{This is underlined text.}
\end{document}
Compiling
texts.tex, as explained in the previous example, will genearte the following output.
Example sections.tex
\documentclass{article}
\begin{document}
\tableofcontents
\section{Animals}
This document denes sections.
\subsection{Domestic}
This document also denes subsections.
\subsubsection{cats and dogs}
Cats and dogs are Domestic animals.
\end{document}
The output of sections.tex is shown in gure 5.1.
2 To
generate the table of contents, you may have to compile the document two times.
CHAPTER 5.
70
5.4 Environments
Environments decide the way in which your text is formatted : numbered lists, tables, equations, quotations, justications, gure, etc. are some of the environments.
Environments are dened like :
\begin{environment_name} your text \end{environment_name}
The example program environ.tex demonstrates some of the environments.
Example environs.tex
\documentclass{article}
\begin{document}
\begin{ushleft} A bulleted list. \end{ushleft}
\begin{itemize} \item dog \item cat \end{itemize}
\begin{center} A numbered List. \end{center}
\begin{enumerate} \item dog \item cat \end{enumerate}
\begin{ushright} This text is right justied. \end{ushright}
\begin{quote}
Any text inside quote\\ environment will appe-\\ ar as typed.\\
\end{quote}
\begin{verbatim}
x=1
while x <= 10:
print x * 5
x=x+1
CHAPTER 5.
71
\end{verbatim}
\end{document}
The enumerate and itemize are used for making numbered and non-numbered
lists. Flushleft, ushright and center are used for specifying text justication. Quote
and verbatim are used for portions where we do not want LATEX to do the formatting.
The output of environs.tex is shown below.
A bulleted list.
dog
cat
A numbered List.
1. dog
2. cat
This text is right justied.
Any text inside quote
environment will appear as typed.
x = 1
#
while x <=
print x
x = x +
a Python program
10:
* 5
1
Example math1.tex
\documentclass{article}
\usepackage{amsmath}
\begin{document}
CHAPTER 5.
72
A B
Am
An
c
ab
ab
3
5
n! = 1 2 (n 1) n
0.3 = 1/3
~a
sin x + arctan y
2
x + y2
z=
x2 +
3 \quad is for inserting space, the size of a \quad corresponds to the width of the character `M' of the current
font. Use \qquad for larger space.
CHAPTER 5.
73
A 6= B
AC
i=1
sinx
2A
x2
b
a
)2
(x + 1)(x 1)
f (x)dx
person
sex
age
John
male
7
M ary f emale 20
Gopal
male
30
The rst column is left justied, second is centered and the third is right justied
(decided by the {|l|cr|}). If you insert a | character between c and r, it will add a
vertical line between second and third columns.
Let us make a matrix using the same command.
$ A = \left(
\begin{array}{ccc}
x_1 & x_2 & \ldots \\
CHAPTER 5.
74
x 1 x2 . . .
y y2 . . .
A=
1
..
.. . .
.
.
.
Learn how to insert pictures with caption inside the gure environment.
20
2.0
15
1.5
10
1.0
0.5
0.0
0.5
10
1.0
15
20
1.5
20
15
10
10
15
20
2.0
Figure 5.2: Picture of Arc and Sawtooth, inserted with [h] option.
CHAPTER 5.
Example qpaper.tex
\documentclass{article}
\usepackage{amsmath}
begin{document}
\begin{center}
\large{\textbf{Sample Question Paper\\for\\
Mathematics using Python}}
\end{center}
\begin{tabular}{p{8cm}r}
\textbf{Duration:3 Hrs} & \textbf{30 weightage}
\end{tabular}\\
\section{Answer all Questions. $4\times 1\frac{1}{2}$}
\begin{enumerate}
\item What are the main document classes in LATEX.
\item Typeset $\sin^{2}x+\cos^{2}x=1$ using LATEX.
\item Plot a circle using the polar() function.
\item Write code to print all perfect cubes upto 2000.
\end{enumerate}
\section{Answer any two Questions. $3\times 5$}
\begin{enumerate}
\item Set a sample question paper using LATEX.
\item Using Python calculate the GCD of two numbers
\item Write a program with a Canvas and a circle.
\end{enumerate}
\begin{center}\text{End}\end{center}
\end{document}
The formatted output is shown below.
75
CHAPTER 5.
76
5.9 Exercises
1. What are the main document classes supported by LATEX.
2. How does Latex dier from other word processor programs.
3. Write a .tex le to typeset 'All types of Text Available' in tiny, large, underline
and italic.
4. Rewrite the previous example to make the output a list of numbered lines.
5. Generate an article with section and subsections with table of contents.
6. Typeset 'All types of justications' to print it three times; left, right and centered.
7. Write a .tex le that prints 12345 in ve lines (one character per line).
8. Typeset a Python program to generate the multiplication table of 5, using
verbatim.
9. Typeset sin2 x + cos2 x = 1
CHAPTER 5.
10. Typeset
11. Typeset
12. Typeset
13. Typeset
x2 + y 2
n=1
A
x
)2
1+
1
n
= x2 + y 2
)n
=A
cos x.dx
14. Typeset x =
b b2 4ac
2a
15. Typeset A =
1 2
3 4
16. Typeset R =
sin cos
cos sin
77
Chapter 6
Numerical methods
Solving mathematical equations is an important requirement for various branches
of science but many of them evade an analytic solution. The eld of numerical
analysis explores the techniques that give approximate but accurate solutions to
such problems.1 Even when they have a solution, for all practical purposes we need
to evaluate the numeric value of the result, with the desired accuracy. We will
focus on developing simple working programs rather than going into the theoretical
details. The mathematical equations giving numerical solutions will be explored by
changing various parameters and nature of input data.
) f (x 4x
)
lim f (x + 4x
2
2
(6.1)
x 0
4x
neglecting the higher order terms. The accuracy of the derivative calculated using
discrete values depends on the stepsize 4x. It will also depends on the number of
higher order derivatives the function has. We will try to explore these aspects
using the program di.py , which evaluates the derivatives of few functions using
two dierent stepsizes (0.1 ans 0.01). The input values to function deriv() are the
function to be dierentiated, the point at which the derivative is to be found and
the stepsize 4x.
Example di.py
def f1(x):
return x**2
1 Introductory methods of numerical analysis by S.S.Sastry
http://ads.harvard.edu/books/1990fnmd.book/
78
CHAPTER 6.
79
NUMERICAL METHODS
4x = 0.005
(b) for
4x = 1.0
def f2(x):
return x**4
def f3(x):
return x**10
def deriv(func, x, dx=0.1):
df = func(x+dx/2)-func(x-dx/2)
return df/dx
print deriv(f1, 1.0), deriv(f1, 1.0, 0.01)
print deriv(f2, 1.0), deriv(f2, 1.0, 0.01)
print deriv(f3, 1.0), deriv(f3, 1.0, 0.01)
The output of the program is shown below. Comparing the two numbers on the
same line shows the eect of stepsize. Comparing the rst number on each row
shows the eect of the number of higher order derivatives the function has. For the
same stepsize x4 gives much less error than x10 . Second derivative of x2 is constant
and the result becomes exact, result on the rst line.
2.0 2.0
4.01 4.0001
10.3015768754 10.0030001575
You may explore other functions by modifying the program. It can be seen that
the function deriv(), evaluates the function at two points to calculate the derivative.
The higher order terms can be calculated by evaluating the function at more points.
Techniques used for this will be discussed in section 6.8, on interpolation.
CHAPTER 6.
NUMERICAL METHODS
80
Figure 6.2: Area under the curve is divided it in to a large number of intervals. Area of each of
them is calculated by assuming them to be trapezoids.
our deriv() function. The dened function is sine and the derivative is calculated
using the vectorized version of deriv(). The actual cosine function also is plotted for
comparison. The output of vdi.py is shown in 6.1(a).
The value of 4x is increased to 1.0 by changing one line of code as y = vecderiv(x, 1.0)
and the result is shown in 6.1(b). The values calculated using our function is shown
using +marker, while the continuous curve is the expected result , ie. the cosine
curve.
Example vdi.py
from pylab import *
def f(x):
return sin(x)
def deriv(x,dx=0.005):
df = f(x+dx/2)-f(x-dx/2)
return df/dx
vecderiv = vectorize(deriv)
x = linspace(-2*pi, 2*pi, 200)
y = vecderiv(x)
plot(x,y,'+')
plot(x,cos(x))
show()
CHAPTER 6.
NUMERICAL METHODS
81
h
[y0 + 2(y1 + y2 + . . . + yn1 ) + yn ]
(6.2)
2
a
This is the sum of the areas of the individual trapezoids. The error in using the
trapezoid rule is approximately proportional to 1/n2 . If the number of sub-intervals
is doubled, the error is reduced by a factor of 4. The program trapez.py does
integration of a given function using equation 6.2. We will choose an example where
the results can be cross checked easily, the value of is calculated by evaluating the
area of a unit circle by integrating the equation of a circle.
ydx =
Example trapez.py
from math import *
def y(x): # equation of a circle
return sqrt(1.0 - x**2)
def trapez(f, a, b, n):
h = (b-a) / n
sum = 0
x = 0.5 * h # f(x) at middle of the slice
for i in range (1,n):
sum = sum + h * f(x)
x = x + h
return sum
print 4 * trapez(y, 0.0, 1.0,1000)
print 4 * trapez(y, 0.0, 1.0,10000)
print trapez(sin,0,2,1000) # Why the error ?
The output is shown below. The result gets better by increasing n thus resulting in
smaller h. The last line shows, how things can go wrong if the arguments are given
in the integer format. Learn how to avoid such pitfalls while writing programs. It
is left as an exercise to the reader to modify the function trapez() to accept integer
arguments also.
3.14041703178
3.14155546691
0.0
CHAPTER 6.
82
NUMERICAL METHODS
dy
= f (x, y); y(x0 ) = y0
(6.3)
dx
where the derivative of the function f (x, y) is known and the value of the function
at some value of x = x0 also is known. The objective is to nd out the value of the
function for other values of x. The underlying idea of any routine for solving the
initial value problem is to rewrite the dy and dx as nite steps 4y and 4x, and
multiply the equations by 4x. This gives algebraic formulas for the change in the
value of y(x) when x is changed by one stepsize 4x . In the limit of making the
stepsize very small, a good approximation to the underlying dierential equation is
achieved.
Implementation of this procedure results in the Euler's method, which is conceptually very important, but not recommended for any practical use. In this section we
will discuss Euler's method and the Runge-Kutta method with the help of example
programs. For detailed information refer to [10, 11].
y (xn , yn ) =
(6.4)
(6.5)
The above equations implies that, if the value of the function y(x) is known to
0
be yn at the point xn , its value at a nearby point xn+1 is given by yn + h y .
The program euler.py calculates the value of sine function using its derivative, ie.
the cosine function. We start from x = 0 , where sin(x) = 0 and compute the
subsequent values using the derivative, cos(x), and compare the result with the
actual sine function.
Example euler.py
from pylab import *
h = 0.01
# stepsize
CHAPTER 6.
NUMERICAL METHODS
83
x = 0.0
# initial values
y = 0.0
ax = [] # Lists to store x and y
ay = []
while x < 2*pi:
y = y + h * math.cos(x) # Euler equation
x = x + h
ax.append(x)
ay.append(y)
plot(ax,ay)
show()
The output of
CHAPTER 6.
NUMERICAL METHODS
84
k1 = hf (xn , yn )
k2 = hf (xn + h2 , yn + k21 )
k3 = hf (xn + h2 , yn + k22 )
k4 = hf (xn + h, yn + k3 )
1
(k1 + 2k2 + 2k3 + k4 )
(6.6)
6
The program rk4.py listed below uses the equations shown above to calculate the
sine function, by integrating the cosine. The output is shown in gure 6.4(a).
yn+1 = yn +
Example rk4.py
from pylab import *
def rk4(x,
k1 = h
k2 = h
k3 = h
k4 = h
return
h = 0.01
# stepsize
x = 0.0
# initial values
y = 0.0
ax = [x]
ay = [y]
while x < math.pi:
y = rk4(x,y,math.cos)
x = x + h
ax.append(x)
ay.append(y)
CHAPTER 6.
NUMERICAL METHODS
85
plot(ax,ay)
show()
The program compareEuRK4.py calculates the values of Sine by integrating Cosine.
The errors in both cases are evaluated at every step, by comparing it with the
sine function, and plotted as shown in gure 6.4(b). The accuracy of Runge-Kutta
method is far superior to that of Euler's method, for the same step size.
Example compareEuRK4.py
from scipy import *
def rk4(x,
k1 = h
k2 = h
k3 = h
k4 = h
return
h = 0.1
# stepsize
x = 0.0
# initial values
ye = 0.0 # for Euler
yr = 0.0 # for RK4
ax = []
# Lists to store results
euerr = []
rkerr = []
while x < 2*pi:
ye = ye + h * math.cos(x) # Euler method
yr = rk4(x, yr, cos, h)
# RK4 method
x = x + h
ax.append(x)
euerr.append(ye - sin(x))
rkerr.append(yr - sin(x))
plot(ax,euerr,'o')
plot(ax, rkerr,'+')
show()
CHAPTER 6.
NUMERICAL METHODS
86
Example rk4_proper.py
def f1(x,y):
return 1 + y**2
def f2(x,y):
return (y-x)/(y+x)
def rk4(x, y, fxy, h): # x, y , f(x,y), step
k1 = h * fxy(x, y)
k2 = h * fxy(x + h/2.0, y+k1/2)
k3 = h * fxy(x + h/2.0, y+k2/2)
k4 = h * fxy(x + h, y+k3)
ny = y + ( k1/6 + k2/3 + k3/3 + k4/6 )
#print x,y,k1,k2,k3,k4, ny
return ny
h = 0.2 # stepsize
x = 0.0 # initial values
y = 0.0
print rk4(x,y, f1, h)
h = 1
x = 0.0
# initial values
y = 1.0
print rk4(x,y,f2,h)
The results are shown below.
0.202707408081
1.5056022409
6.4 Polynomials
A polynomial is a mathematical expression involving a sum of powers in one or more
variables multiplied by coecients. A polynomial in one variable with constant
coecients is given by
an xn + ... + a2 x2 + a1 x + a0
(6.7)
The derivative of 6.7 is,
CHAPTER 6.
NUMERICAL METHODS
87
Example poly.py
from pylab import *
a = poly1d([3,4,5])
b = poly1d([6,7])
c = a * b + 5
d = c/a
print a
print a(0.5)
print b
print a * b
print a.deriv()
print a.integ()
print d[0], d[1]
The output of
CHAPTER 6.
88
NUMERICAL METHODS
3x2 + 4x + 5
7.75
6x + 7
18x3 + 45x2 + 58x + 35
6x + 4
1x3 + 2x2 + 5x
6x + 7
5
The last two lines show the result of the polynomial division, quotient and reminder. Note that a polynomial can take an array argument for evaluation to return
the results in an array. The program polyplot.py evaluates polynomial 6.8 and its
rst derivative.
x3
x5
x7
(6.8)
x
+
6
120 5040
The results are shown in gure 6.5. The equation 6.8 is the rst four terms
of series representing sine wave (7! = 5040). The derivative looks like cosine as
expected. Try adding more terms and change the limits to see the eects.
Example polyplot.py
from pylab import *
x = linspace(-pi, pi, 100)
a = poly1d([-1.0/5040,0,1.0/120,0,-1.0/6,0,1,0])
da = a.deriv()
y = a(x)
y1 = da(x)
plot(x,y)
plot(x,y1)
show()
(x a)2 00
(x a)n n
f (x) = f (a) + (x a)f (a) +
f (a) + +
f (a)
2!
n!
For example let us consider the equation
0
(6.10)
f (x) = x3 + x2 + x
We can see that f (0) = 0 and the derivatives are
f 0 (x) = 3x2 + 2x + 1;
f 00 (x) = 6x + 2;
(6.9)
f 000 (x) = 6
CHAPTER 6.
89
NUMERICAL METHODS
Using the values of the derivatives at x = 0and equation 6.9, we evaluate the
function at x = .5, using the polynomial expression,
f (0.5) = 0 + 0.5 1 +
0.52 2 0.53 6
+
= .875
2!
3!
The result is same as 0.53 + 0.52 + 0.5 = .875. We have calculated it manually
for x = .5 . We can also do this using Numpy as shown in the program taylor.py.
Example taylor.py
from numpy import *
p = poly1d([1,1,1,0])
dp = p.deriv()
dp2 = dp.deriv()
dp3 = dp2.deriv()
a = 0 # The known point
x = 0 # Evaluate at x
while x < .5:
tay = p(a) + (x-a)* dp(a) + \
(x-a)**2 * dp2(a) / 2 + (x-a)**3 * dp3(a)/6
print '%5.1f %8.5f\t%8.5f'%(x, p(x), tay)
x = x + .1
The result is shown below.
0.0
0.1
0.2
0.3
0.4
0.00000
0.11100
0.24800
0.41700
0.62400
0.00000
0.11100
0.24800
0.41700
0.62400
sin(x) = x
x 3 x5
+
+
3!
5!
n=0
(1)n
x2n+1
(2n + 1)!
(6.11)
CHAPTER 6.
90
NUMERICAL METHODS
cos(x) = 1
x2 x 4
+
+
2!
4!
n=0
(1)n
x2n
(2n)!
(6.12)
The program series_sc.py evaluates the sine and cosine series. The output is
shown in gure 6.6. Compare the output of polyplot.py from section 6.4 with this.
In both cases, we have evaluated the polynomial of sine function. In the present
case, we can easily modify the number of terms and the logic is simpler.
Example series_sc.py
from pylab import *
def f(n):
# Factorial function
if n == 0:
return 1
else:
return n * f(n-1)
NP = 100
x = linspace(-pi, pi, NP)
sinx = zeros(NP)
cosx = zeros(NP)
for n in range(10):
sinx += (-1)**(n) * (x**(2*n+1)) / f(2*n+1)
cosx += (-1)**(n) * (x**(2*n)) / f(2*n)
plot(x, sinx)
plot(x, cosx,'r')
show()
CHAPTER 6.
NUMERICAL METHODS
91
Example rootsearch.py
def func(x):
return x**3-10.0*x*x + 5
def root(f,a,b,dx):
x = a
while True:
f1 = f(x)
f2 = f(x+dx)
if f1*f2 < 0:
return x, x + dx
x = x + dx
if x >= b:
return (None,None)
x,y =
print
x,y =
print
The outputs are (0.7 , 0.8) and (1.5 , 1.6). The root of cosine,/2, is between 1.5
and 1.6. After the root has been located roughly, we can nd the root with any
specied accuracy, using bisection method, Newton-Raphson method etc.
CHAPTER 6.
92
NUMERICAL METHODS
2x2 3x 5
x=4
and
x=4
(b) tan(x).
Example bisection.py
import math def func(x):
return x**3 - 10.0* x*x + 5
def bisect(f, x1, x2, epsilon=1.0e-9):
f1 = f(x1)
f2 = f(x2)
if f1*f2 > 0.0:
print 'x1 and x2 are on the same side of x-axis'
return
n = math.ceil(math.log(abs(x2 - x1)/epsilon)/math.log(2.0))
CHAPTER 6.
93
NUMERICAL METHODS
n = int(n)
for i in range(n):
x3 = 0.5 * (x1 + x2)
f3 = f(x3)
if f3 == 0.0: return x3
if f2*f3 < 0.0:
x1 = x3
f1 = f3
else:
x2 = x3
f2 = f3
return (x1 + x2)/2.0
print bisect(func, 0.70, 0.8, 1.0e-4)
print bisect(func, 0.70, 0.8, 1.0e-9)
xi+1 = xi
f (xi )
f 0 (xi )
(6.14)
Figure 6.7(a) shows the graph of the quadratic equation 2x2 3x 5 = 0 and its
two tangents. It can be seen that the zeros are at x = -1 and x = 2.5, and we use
the program newraph.py shown below to nd the roots. The function nr() is called
twice, and we get the roots nearer to the corresponding starting values.
Example newraph.py
from pylab import *
def f(x):
return 2.0 * x**2 - 3*x - 5
def df(x):
return 4.0 * x - 3
def nr(x, tol = 1.0e-9):
CHAPTER 6.
NUMERICAL METHODS
94
for i in range(30):
dx = -f(x)/df(x)
#print x
x = x + dx
if abs(dx) < tol:
return x
print nr(4)
print nr(0)
The output is shown below.
2.5
-1.0
Uncomment the print statement inside nr() to view how fast this method converges, compared to the bisection method. The program newraph_plot.py, listed
below is used for generating the gure 6.7.
Example newraph_plot.py
from pylab import *
def f(x):
return 2.0 * x**2 - 3*x - 5
def df(x):
return 4.0 * x - 3
vf = vectorize(f)
x = linspace(-2, 5, 100)
y = vf(x)
# Tangents at x=3 and 4, using one point slope formula
x1 = 4
tg1 = df(x1)*(x-x1) + f(x1)
x1 = 3
tg2 = df(x1)*(x-x1) + f(x1)
grid(True)
plot(x,y)
plot(x,tg1)
plot(x,tg2)
ylim([-20,40])
show()
We have dened the function f (x) = 2x2 3x 5 and vectorized it. The derivative
4x2 3 also is dened by df (x), which is the slope of f (x). The tangents are drawn
at x = 4 and x = 3, using the point slope formula for a line y = m(x x1) + y1.
CHAPTER 6.
95
NUMERICAL METHODS
4x + y 2z = 0
2x 3y + 3z = 9
6x 2y + z = 0
can be represented in the matrix form as
4
1 2
x
0
y
2
3
3
= 9
6 2
1
z
0
and can be solved by nding the inverse of the coecient matrix.
x
4
1 2
3
y = 2 3
z
6 2
1
9
0
CHAPTER 6.
96
NUMERICAL METHODS
S(a0 , a1 , . . . , am ) =
(6.15)
i=0
S
= 0,
ai
For a linear t, the equation is
i = 0, 1, . . . m
(6.16)
f (a, b) = a + bx
Solving the equations
S
a
= 0 and
b=
S
b
yi (x x)
, and a = y xb
xi (x x)
(6.17)
x=
n
1
xi ,
n + 1 i=0
y=
n
1
yi
n + 1 i=0
The program lst.py demonstrates the usage of equations 6.17 and 6.18.
Example lst.py
(6.18)
CHAPTER 6.
97
NUMERICAL METHODS
6.8 Interpolation
Interpolation is the process of constructing a function f (x) from a set of data points
(xi , yi ), in the interval a < x < b that will satisfy yi = f (xi ) for any point in the
same interval. The easiest way is to construct a polynomial of degree n that passes
through the n + 1 distinct data points.
(6.20)
yo = a0
For x = x1 , only the rst two terms will be non-zero.
(6.21)
CHAPTER 6.
98
NUMERICAL METHODS
y1 = a0 + a1 (x1 x0 )
a1 =
(6.22)
(y1 y0 )
(x1 x0 )
(6.23)
Applying x = x2 , we get
(6.24)
0
xy11 y
x0
(6.25)
x2 x0
The other coecients can be found in a similar manner. They can be expressed
better using the divided dierence notation as shown below.
a2 =
y2 y1
x2 x1
[y0 ] = y0
[y0 , y1 ] =
[y0 , y1 , y3 ] =
(y1 y0 )
(x1 x0 )
y2 y1
x2 x1
0
xy11 y
[y1 , y2 ] [y0 , y1 ]
x0
=
x 2 x0
(x2 x0 )
(6.26)
The divided dierence can be put in the tabular form as shown below. This will
be useful while calculating the coecients manually.
x0
x1
x2
x3
x0
y1
y2
y3
0
1
2
3
[y0 ]
[y1 ] [y0 , y1 ]
[y2 ] [y1 , y2 ] [y0 , y1 , y2 ]
[y3 ] [y2 , y3 ] [y1 , y2 , y3 ] [y0 , y1 , y2 , y3 ]
0
3
14
39
30
10
143
21
3914
32
=3
= 11
= 25
113
20
2511
31
=4
=7
74
30
=1
The table given above shows the divided dierence table for the data set x =
[0,1,2,3] and y = [0,3,14,39], calculated manually. The program newpoly.py can be
used for calculating the coecients, which prints the output [0, 3, 4, 1].
CHAPTER 6.
NUMERICAL METHODS
99
Example newpoly.py
from copy import copy
def coef(x,y):
a = copy(y)
m = len(x)
for k in range(1,m):
tmp = copy(a)
for i in range(k,m):
tmp[i] = (a[i] - a[i-1])/(x[i]-x[i-k])
a = copy(tmp)
return a
x = [0,1,2,3]
y = [0,3,14,39]
print coef(x,y)
We start by copying the list y to coecient a, the rst element a0 = y0 . While
calculating the dierences, we have used two loops and a temporary list. The same
can be done in a better way using arrays of Numpy2 , as shown in newpoly2.py.
Example newpoly2.py
from numpy import *
def coef(x,y):
a = copy(y)
m = len(x)
for k in range(1,m):
a[k:m] = (a[k:m] - a[k-1])/(x[k:m]-x[k-1])
return a
x = array([0,1,2,3])
y = array([0,3,14,39])
print coef(x,y)
The next step is to calculate the value of y for any given value of x, using the coecients already calculated. The program newinterpol.py calculates the coecients
using the four data points. The function eval() uses the recurrence relation
(6.27)
function is from reference [12], some PDF versions of this book are available on the web.
CHAPTER 6.
NUMERICAL METHODS
Example newinterpol.py
from pylab import *
def eval(a,xpoints,x):
n = len(xpoints) - 1
p = a[n]
for k in range(1,n+1):
p = a[n-k] + (x -xpoints[n-k]) * p
return p
def coef(x,y):
a = copy(y)
m = len(x)
for k in range(1,m):
a[k:m] = (a[k:m] - a[k-1])/(x[k:m]-x[k-1])
return a
x = array([0,1,2,3])
y = array([0,3,14,39])
coef = coef(x,y)
NP = 20
newx = linspace(0,3, NP) # New x-values
newy = zeros(NP)
for i in range(NP): # evaluate y-values
newy[i] = eval(coef, x, newx[i])
plot(newx, newy,'-x')
plot(x, y,'ro')
show()
100
CHAPTER 6.
NUMERICAL METHODS
101
You may explore the results for new points outside the range by changing the second
argument of line newx = linspace(0,3,NP) to a higher value.
Look for similarities between Taylor's series discussed in section 6.4.1 that and
polynomial interpolation process. The derivative of a function represents an innitesimal change in the function with respect to one of its variables. The nite difference is the discrete analog of the derivative. Using the divided dierence method,
we are in fact calculating the derivatives in the discrete form.
6.9 Exercises
1. Dierentiate 5x2 + 3x + 5 numerically and evaluate at x = 2 and x = 2.
2. Write code to numerically dierentiate sin(x2 ) and plot it by vectorizing the
function. Compare with the analytical result.
3. Integrate ln x, ex from x = 1to 2.
4. Solve 2x + y = 3; x + 4y = 0; 3 + 3y = 1 using matrices.
5. Modify the program julia.py, c = 0.2 0.8j and z = z 6 + c
6. Write Python code, using pylab, to solve the following equations using matrices
4x + y 2z = 0
2x 3y + 3z = 9
6x 2y + z = 0
7. Find the roots of 5x2 + 3x 6 using bisection method.
8. Find the all the roots of sin(x) between 0 and 10, using Newton-Raphson
method.
The System
This section will describe how Ubuntu was installed on a system, with MSWindows, having the following partitions:
C: (GNU/Linux calls it /dev/sda1) 20 GB
D: ( /dev/sda5) 20 GB
E: (/dev/sda6) 30 GB
We will use the partition E: to install Ubuntu, it will be formatted.
The Procedure
Set the rst boot device CD ROM from the BIOS. Boot the PC from the Ubuntu
installation CD, we have used Phoenix Live CD (a modied version on Ubuntu 9.1).
After 2 to 3 minutes a desktop as shown below will appear. Click on the Installer
icon, the window shown next will pop up. Screens will appear to select the language,
time zone and keyboard layout as shown in the gures below.
102
Installing Ubuntu
103
The bar on the top graphically displays the existing partitions. Below that there
are three options provided :
Installing Ubuntu
104
If we proceed with this, a warning will appear complaining about the absence of
swap partitions. The swap partition is used for supplementing the RAM with some
virtual memory. When RAM is full, processes started but not running will be
swapped out. One can install a system without swap partition but it is a good idea
to have one.
Installing Ubuntu
105
We decide to go back on the warning, to delete the E: drive, create two new
partitions in that space and make one of them as swap. This also demonstrates how
to make new partitions. The screen after deleting E: , with the pop-up window to
make the swap partition is shown below.
We made a 2 GB swap. The remaining space is used for making one more
partition, as shown in the gure 6.10.
Once disk partitioning is over, you will be presented with a screen to enter a
user name and password.3 A warning will be issued if the password is less than 8
characters in length. You will be given an option to import desktop settings from
other installations already on the disk, choose this if you like. The next screen will
conrm the installation. After the installation is over, mat take 10 to 15 minutes,
you will be prompted to reboot the system. On rebooting you will be presented
with a menu, to choose the operating system to boot. First item in the menu will
be the newly installed Ubuntu.
106
Installing Ubuntu
enormous number of packages, that can be installed very easily. You need to have
a reasonably fast Internet connection for this purpose.
From the main menu, open System->Administration->Synaptic package manager. After providing the pass word (of the rst user, created during installation),
the synaptic window will popup as shown in gure 6.11.
Select Settings->Repositories to get a pop-up window as shown below. Tick
the four repositories, close the pop-up window and Click on Reload. Synaptic will
now try to download the index les from all these repositories. It may take several
minute.
107
Installing Ubuntu
Now, you are ready to install any package from the Ubuntu repository. Search
Installing Ubuntu
108
for any package by name, from these repositories, and install it. If you have done
the installation from original Ubuntu CD, you may require the following packges:
lyx : A latex front-end. Latex will be installed since lyx dpends on latex.
python-matplotlib : The graphics library
python-visual : 3D graphics
python-imaging-tk : Tkinter, Python Imaging Library etc. will be installed
build-essential : C compiler and related tools.
Installing Ubuntu
109
The storage space of a hard disk drive can be divided into separate data areas,
known as partitions. You can create primary partitions and extended partitions.
Logical drives (secondary partitions can be created inside the extended partitions).
On a disk, you can have up to 4 partitions, where one of them could be an extended
partition. You can have many logical drives inside the extended partition.
On a MSWindows system, the primary partition is called the C: drive. The logical
drives inside the extended partition are named from D: onwards. GNU/Linux uses a
dierent naming convention. The individual disks are named as /dev/sda , /dev/sdb
etc. and the partitions inside them are named as /dev/sda1, /dev/sda2 etc. The
numbering of secondary partitions inside the logical drive starts at /dev/sda5. (1 to
4 are reserved for primary and extended). Hard disk partitioning can be done using
the fdisk program. The installation program also does this for you.
The process of making a le system on a partition is called formatting. There
are many dierent types of le systems. MSWindows use le systems like FAT32,
NTFS etc. and GNU/Linux mostly uses le systems like ext3, ext4 etc.
The operating system les are kept in directories named boot, sbin, bin, etc etc.
The kernel that loads while booting the system is kept in /boot. The conguration
les are kept in /etc. /sbin and /bin holds programs that implements many of the
shell commands. Most of the application programs are kept in /usr/bin area.
The boot loader program is the one provides the selection of OS to boot, when
you power on the system. GRUB is the boot loader used by most of the GNU/Linux
systems.
Bibliography
[2] http://en.wikipedia.org/wiki/List_of_curves
[3] http://www.gap-system.org/~history/Curves/Curves.html
[4] http://www.gap-system.org/~history/Curves/Ellipse.html
[5] http://mathworld.wolfram.com/
[6] http://www.scipy.org/Numpy_Example_List
[7] http://docs.scipy.org/doc/
[8] http://numericalmethods.eng.usf.edu/mws/gen/07int/index.html
[9] http://www.angelre.com/art2/fractals/lesson2.htm
[10] http://www.zyka.umk.pl/nrbook/bookcpdf.html
[11] http://www.mathcs.emory.edu/ccs/ccs315/ccs315/ccs315.html
[12] Numerical Methods in Engineering with Python by Jaan Kiusalaas
110